Crisis Communication Channels: Recommendations

By Matt Abud 2013 ACKNOWLEDGEMENTS Internews and Google’s Disaster Response Team share a deep commitment to developing informa- tion tools that can mediate the challenges associated with humanitarian crises. Tech tools posi- tioned in the hands of the right people can reunite families, connect people to safe havens in the midst of a disaster, or help humanitarian organizations navigate the quickly changing terrain when normal life is disrupted by disaster. In 2012, Google’s Disaster Response team and Internews’ Humanitarian Communication Program began to look at the range and use of tech tools by communities affected by crisis in the developing world. Internews’ Center for Innovation and Learning came forward to contribute a design meth- odology that could help answer basic questions about what tools have been used in three different disasters in Indonesia: Jakarta’s January 2013 floods; Aceh’s April 2012 earthquake and tsunami warning; and the ongoing volcano eruption in Rokatenda, Flores, which began in late 2012. This document tells that story and opens up a broader dialogue of how we can understand the needs of people and organizations in the midst of crisis and develop new protocols, systems, and tools that enhance connection and save lives in times of crisis. The Internews’ Humanitarian Media Team takes a leadership role in understanding how media and communications plays a role in humanitarian crises. Many thanks are due to a great number of people who helped during this research. They include: the Google Crisis Response team, both for supporting the research and their questions and discussion along the way. Gladys Respati and the whole team at OnTrack Media Indonesia, for going beyond the call with all collaboration and support provided. Juni Soehardjo, for research and several insights in the section on national issues. Many others provided both their time, and greatly facilitated further interviews. They include: colleagues at Palang Merah Indonesia and several IFRC member societies operating in Indonesia, including American Red Cross in Banda Aceh. Staff of local disaster manage- ment agencies in Jakarta; in Kupang and Sikka in east Indonesia (BPBD); and in Aceh (BPBA); and of the national disaster management agency (BNPB). Staff at UNOCHA and at the Australia Indonesia Disaster Reduction Facility. The Urban Poor Consortium for facilitating much of the fieldwork in north Jakarta. And of course numerous journalists, editors, humanitarian workers and digital activists in each of the areas researched gave invaluable insights throughout. Most of all, thanks are due to the residents along the Ciliwung River and in Muara Baru in Jakarta; to the residents displaced from Palue Island; and the residents in Banda Aceh who shared their experi- ences and perspectives.

Credits Design: Kirsten Ankers, Citrine Sky Design Front cover photos (from left to right): Oren Murphy, Febi Dwirahmadi and Matt Abud. Back cover photo: Oren Murphy Contents

1. Introduction...... 2 2. Overview...... 3 3. Methodology...... 4 4. Google Services in Disaster Response: Broad Findings...... 6 4.1. Google Approaches to Indonesia...... 6 4.1.1. Google Tool Design...... 6 4.1.2. Business Strategies...... 6 4.1.3 Targeted Audiences, Targeted Criteria, Targeted Input...... 7 4.1.4 Targeted Partnerships...... 7 4.2. Information Sources before the flood: Where Google tools fit...... 8 4.2.1 Information Sources Accessed...... 8 4.2.2 Trust in Sources...... 8 4.3. Information Sources during the flood: Where Google tools fit...... 9 4.3.1 Google Tools accessed...... 9 4.3.2 Rokatenda and Aceh case-studies...... 10 4.3.3 Use of Google Tools in the Jakarta Floods: Some Observations...... 10 5. Google Services in Disaster Response: Case Study Example...... 11 5.1 Disaster Agencies: Jakarta...... 11 5.2 Media: Metro TV, others...... 12 5.3 NGOs & Civil Society: RAPI...... 12 5.4 Coordination: mailing list...... 12 5.5 Digital Volunteers: ODOS...... 13 5.6 Digital volunteers: groups...... 13 6. Potential Initiatives...... 14 6.1 Neighborhood Social Networks...... 15 6.2 Government Agencies...... 16 6.3 Media...... 17 6.4 Telecommunications Sector...... 18 6.5 NGOs & Civil Society...... 19 6.6 Coordination...... 19 6.7 Digital Volunteers & Communities of Practice...... 19 7. List of Acronyms...... 21

Crisis Communication Channels: Google Recommendations 1 1 Introduction This Recommendations document is a specific discussion focused on Google and Google Crisis Response in Indonesia’s disaster responses, accompanying and drawing on the findings of Internews’ broader Crisis Communications Channels report. That report explores communications ecosystems in three recent disaster response case studies.

Overall findings in theseRecommendations show a low level of use of Google and Google Crisis Response tools by both affected communities and, perhaps more surprisingly, respond- ers themselves. The report details the individual examples of Google tool use that were encountered, including a review of possible reasons or dynamics that prevent such use from reaching its full potential. The goal of this research is not to point out limitations, but to suggest potential ways forward. The approach of Crisis Communication Channels, and of this accompanying Recommendations document, rests on an appreciation of the relationships that underpin effective communication. A single tool, service, or piece of information is frequently not enough to provide effective communication. Tools must be familiar and understood; information must be believed; data must flow rap- idly to and from actors who know how to use it best; communi- ties must be heard and receive responses to their concerns. All of these are necessary, with the absence of any one element potentially undermining all. Residents have been forced to adapt to frequent floods, with some closest to the river inundated eight times by late February. Photo Finally, it is pertinent to note that Indonesia was selected as by Oren Murphy. an example of a developing country that nevertheless shows a wide range of circumstances across a broad geographical area. In keeping with the focus on information and communication Many of the features and dynamics identified in the report will as a life-saving resource in a disaster, Internews focused on have resonance with other developing contexts. However fur- the experiences of those most affected and in need, and on ther comparative studies will help refine and isolate those fea- the responders whose efforts could meet those needs. The tures that generally hold true, and others which may be down research does not, therefore, incorporate the experiences of to certain factors — environmental, policy, or otherwise – par- those who were less affected and whose lives and wellbeing ticular to a specific country. were not placed at serious risk, even though they may live in nearby or adjoining areas and therefore be affected to a lesser degree.

2 Crisis Communication Channels: Google Recommendations Overview

These Recommendations draw on the far more detailed case studies in the Crisis Communications Channels2 report. Specific sections in that report are referenced when needed, rather than repeating large slabs of the same detail here.

The Recommendations document first presents the methodol- ogy used to gather this data. It then discusses overall findings Government Agencies on the use of Google Tools, by disaster-affected populations, n Media (meaning conventional media outlets: T.V., radio, and by responders and other institutional actors. It then pres- newspapers) ents detailed individual examples of how Google tools were n Telecommunications sector used, drawn from the Crisis Communications Channels report n NGOs and Civil Society organisations and highlighting a range of actors. Finally, it outlines a series n Coordination (while not an ‘actor’ this cross-cutting area is of potential initiatives that can support improved use of Google highlighted in recognition of the key role it plays in facilitat- tools for the needs of specific actors. ing an effective response, including effective communica- The actors included in this discussion are the same as those in tion flows) the Crisis Communication Channels, namely: n Digital Volunteers and Communities of Practice (that Neighbourhood Social Networks (networks among affected is, networks connected over and mobilised through digital populations, largely generated through face-to-face or proxim- platforms to provide volunteer humanitarian response) ity of contact)

Church delivery to aid post in Ende regency, Flores. Photo by Matt Abud / Internews

Crisis Communication Channels: Google Recommendations 3 Methodology 3 Crisis Communications Channels rests on three disaster response case studies: Jakarta’s January 2013 floods; Aceh’s April 2012 earthquake and tsunami warning; and the ongoing volcano eruption in Rokatenda, Flores, which began in late 2012. Case studies were selected to provide a wide range of contexts, and included loca- tions that are rich in communication infrastructure (Jakarta) and poor in infrastructure (Rokatenda); sudden early-warning scenarios (Aceh) and both long- and short-term disasters (Rokatenda and Jakarta); and dif- ferent disaster types (earthquake and tsunami; volcanic eruption; and urban flood), among other variations.

disaster types (earthquake and tsunami; volcanic eruption; and urban flood), among other variations. Internews conducted over one hundred key interviews in the course of gathering case study material. These included inter- views with affected communities (approximately 36 people);1 Local Government and Government Disaster Agency staff (10); media, both journalists and editors (20); telecommunications company staff (3); NGOs and civil society (27); and digital vol- unteers (11). In each interview Internews asked several questions about both the use of online tools in general, and — if the respon- dent had any online engagement at all — Google and Google Crisis Response tools in particular. These were open questions, to ensure any tool utilised would be mentioned. Follow-up questions included asking whether of any sort were used; and subsequently whether Google Crisis Maps were used specifically. Aceh 2004 tsunami damage. Photo by Oren Murphy. In addition, Internews independently funded in-depth quan- titative surveys of selected flood-affected communities in Crisis Communications Channels rests on three disaster Jakarta to contextualise and triangulate findings from the key response case studies: Jakarta’s January 2013 floods; Aceh’s interviews. 300 respondents answered questions both on their April 2012 earthquake and tsunami warning; and the ongo- regularly accessed channels of information pre-flood, and the ing volcano eruption in Rokatenda, Flores, which began in late channels of information accessed during the flood. Questions 2012. Case studies were selected to provide a wide range of placed online services, including those provided by Google, in contexts, and included locations that are rich in communication the context of broader communication channels. Jakarta was infrastructure (Jakarta) and poor in infrastructure (Rokatenda); targeted for more in-depth surveys as the disaster response

sudden early-warning scenarios (Aceh) and both long- and 1 Some interviews were with small groups, with input from both a group short-term disasters (Rokatenda and Jakarta); and different representative and some detailed contributions from group members; the actual number of contributors is therefore higher than this

4 Crisis Communication Channels: Google Recommendations had the most involved communication channels, including significant penetration of online services; and because Google Crisis Response specifically provided services towards this response effort. The sample size is adequate to draw informed conclusions and make subsequent recommendations.2

2 Internews contracted an independent research organisation to conduct surveys on media and information access among flood-affected communities. This targeted the same communities along the Ciliwung River and in Muara Baru visited by Internews’ researchers themselves, in order to correlate information from those qualitative interviews with quantitative data. In addition for further comparison the survey covered three other locations: broader areas of Pluit (of which Muara Baru is a part); and Rawa Terate and Kampung Melayu, both in East Jakarta.

The survey sample size was 300, high for the size of the target population and giving a level of reliability to the results. The sample targeted citizens 15 years old and over; in order to adequately represent younger demographics whom interviews indicated had different information consumption habits, especially using online platforms, 10% of all respondents were between 15 and 18 years old. Respondents were purposively selected to target the flood- affected, as flood impact was highly-localised even within individual neighborhoods. This meant that randomised selection of households would not necessarily yield an accurate sample, as households within a short distance of each other may not have all been flood affected. 90% of respondents were purposively selected to target flood-affected households, with demographics randomized using the Kish grid methodology. A further 10% were purposively selected to target those who were not flood-affected themselves, but resided nearby and assisted those who were affected. This attempted to draw out differences in information access caused by diverse individual circumstances that were nevertheless in close proximity. Given the fact that this study was conducted using non-probability convenience sampling methods, the results are not projectable to any population beyond the respondents themselves.

Interviewers received training and briefings from the research company, with Internews in attendance at the start-up for any orientation and clarification. Interviews were conducted face-to-face, with results recorded on paper and subsequently entered into the data-base. 130 of those approached for interviews refused, at a rate of 20%, with the most common reason being lack of time to participate.

The team field coordinator witnessed 10% of interviews on-site. Call- back for quality control covered a further 20% of the total respondents, completed by phone. Quality control thus covered 30% of total interviews. In the event of any error discovered, the questionnaires of the interviewer concerned were then double-checked.

The survey targeted three main areas: information sources regularly accessed before the floods; information sources accessible during the floods, to provide a comparison of how sources changed or were impacted by the disaster; and which sources were useful in accessing assistance or making decisions. Throughout, questions also focused on influences that were changing The survey targeted three main areas: information sources regularly accessed before the floods; information sources accessible during the floods, to provide a comparison of how sources changed or were impacted by the disaster; and which sources were useful in accessing assistance or making decisions. Throughout, questions also focused on influences that were changing communication in some detail, in particular the use of mobile phones and online platforms.

Crisis Communication Channels: Google Recommendations 5 Google Services in Disaster Response: Broad Findings 4From the key interviews, the almost universal feedback indicates that, in general, Google Crisis Response tools were either under-utilised or not well-known among public and local crisis response practitioners in Indonesia.

4.1. Google Approaches to Indonesia The suggested initiatives to potentially improve this practice, outlined at the end of these Recommendations, largely rest on developing partnerships with key actors, with the purpose of aligning Google Crisis Response tool, expertise, and resources with other specialist roles and skills others hold in a response effort. Internews recognizes that the practicality any initiative depends upon how it fits with the core roles and ambitions Google Crisis Response defines for itself; this includes whether a greater number of partnerships ‘on the ground’ in a response is desirable or even operationally feasible. However some initiatives also rest on the approaches and resources Google Crisis Response, and Google itself, have in developing tools for markets like that of Indonesia. These fac- tors, and accompanying potential initiatives, are noted here. Informal networks play an essential role in emergency communi- They include: cation flows — especially when conventional media is not geared to meeting these local needs. But even in information-rich Jakarta, informal networks in marginalized communities are often forced to 4.1.1. Google Tool Design take up a greater role than they can handle effectively. Photo by The design of some Google Crisis Response tools appears to Febi Dwirahmadi. hamper their utility in disasters in developing contexts such as represented that represented by Indonesia. In particular, Google tools were sometimes used, but only occasionally and in Internews’ own experience the Crisis Response Map is opti- incidentally for specific response purposes. No citizens, respond- mised for desktop browsers (as is the Crisis Response website ers, or media reported using either the MyMap or the Crisis itself). But Indonesian citizens, and especially those most Response Map to get flood status updates; no responders report- vulnerable, rely heavily on mobile connectivity. Potential ini- ing using the Maps to help plan or organise their response. The tiatives include optimizing Crisis Maps for mobile use, or — if exception was an extremely limited use of the general Google practical from an engineering point of view, an assessment of Maps platform by ODOS (One Day One Service) volunteers (see which is beyond Internews’ expertise — as layers for mobile Section 3.11 in Crisis Communications Channels). Google Map apps.

6 Crisis Communication Channels: Google Recommendations 4.1.2. Business Strategies practical outcomes inasmuch as some developer informants Google’s business structures and approach in Indonesia were unsure of what costs may be involved in designing or emerged as a significant influence in the development of deploying certain applications, and how to find out those costs. digital crisis communications practice, as supported by Google Greater clarity of Google’s approach, and of potential collabora- Crisis Response. Internews does not have broad knowledge tion, utilisation, or co-leverage of Google Crisis Response, may of Google’s overall approach to the Indonesian market. Some assist in boosting involvement of key digital humanitarian ini- of the issues raised are noted here, as they may help inform tiatives within the country. Google’s internal deliberations. 4.1.3 Targeted Audiences, Targeted Criteria, Service Charges: data bases. Jakarta’s Local Disaster Management Agency (BPBD, Badan Penanggulangan Bencana Targeted Input Google Crisis Response supports tools targeting different Daerah) informed Internews that it was unable to use Google actors and audiences, ranging from data base services for services for data management because they were too expen- responders, to crisis maps for both responders and the pub- sive. While this may require confirmation, if it is the case, it lic. Individuals and organizations that did use or access tools is possible that some of Google’s overall business decisions happily gave feedback. However it is difficult to measure this or imperatives run counter to some of its humanitarian crisis feedback beyond individual assessments of efficacy. A clear set response goals. of criteria from Google Crisis Response itself, shared and pub- It may be possible to calibrate the two – for example, by grant- licised with the targeted audience for a given tool or service, ing humanitarian exceptions for the use of some needed data could help evaluate effectiveness and point towards future services by specific institutions, or through other arrangements. improvements. The very development of such a set of crite- Service Charges: Mobile Apps. Humanitarian programmers ria, if it draws in representatives of target audiences, can itself and independent mobile app developers also cite cost for avoid- clarify further positive directions. ing the use of Google Maps in their app development; several are turning to Open Street Maps instead. Humanitarian apps 4.1.4 Targeted Partnerships aim for high use at no cost to the user; however as explained to As Crisis Communication Channels documents, and these Internews, if the number of users increases, charges incur for Recommendations refer to, communication flows in a cri- the use of Google Map data. Again, this may be an example of sis incorporate a wide range of dynamics. Effective ‘last mile’ business imperatives hampering potential humanitarian appli- communications, particularly in developing contexts such as cations. The detail of these charges, or whether calibration Indonesia, must take account of the perceptions and concerns or exceptions for humanitarian applications approach is even of local communities; that is, they must inevitably draw on technically practical, is beyond Internews’ area of expertise; community development approaches. Effective communica- however the feedback from local developers was consistent on tions flows within key institutions — such as disaster response this point. agencies — are inevitably caught up in those same institutions’ development trajectory. Reporting by journalists is affected by . A further barrier to mobile app development and maturity of the country’s media sector; and so on. dissemination targeting Indonesia’s humanitarian needs is the fact that Indonesian-based credit cards are not accepted on Google and Google Crisis Response naturally would not aim to Google Play. Independent app developers are therefore limited become community development, institutional development, to providing apps to Indonesian companies, for delivery via or media development specialists. A key question, therefore, those companies’ own websites. This inhibits integration with is what kinds of partnerships are feasible in order to connect and adoption through Google’s own ecosystem, and fragments Google Crisis Response to those skills and approaches that can and restricts distribution capabilities. Again, Internews is not enhance ongoing development and deployment of its tools in aware of the business reasons underlying this situation. targeted areas. Many of the initiatives suggested at the end of these Recommendations rest on developing such partnerships. Access and Awareness. Although the wide range of thematic Those initiatives are exactly that – suggestions – and naturally areas required by Crisis Communications Channels meant need to be balanced against the broad range of priorities and Internews only met a limited number of developers, several programs that Google Crisis Response carries out. indicated confusion or lack of clarity on decisions surrounding Google’s approach to the Indonesian market — for example, why it did not have an office in country, and so on. This had

Crisis Communication Channels: Google Recommendations 7 this question on pre-flood information sources, which feedback 4.2. Information Sources before from face-to-face interviews indicated were the most widely- the flood: Where Google tools fit used tools in the target context.

4.2.1 Information Sources Accessed 4.2.2 Trust in Sources From the Jakarta survey of affected communities, around 14% Trust in online tools is mixed, likely reflecting unfamiliarity in of respondents indicated they used on a regular their use by most users who are on the wrong side of the digital basis, with a tendency both towards youth and male users. divide. Trust in Google tools reflects this. This may be due to (Facebook, despite impressive overall figures within Indonesia, the fact that digital sources are new, and other sources — T.V., registered lower at 11%.) Google Maps rated 2% within the local government, family and friends — are well-established target population. These were the two Google tools included in and their nature more ‘understood’.

Chart 1: Important Sources of Information Q8: What are important sources of information for you? Select as many answers as relevant.

Chart 2: Trust in Information Sources

Q11: Now I would like to ask how much you trust the news and information you hear from different source. From what you know, is the news and information on………very trustworthy, somewhat trustworthy, somewhat untrust- worthy or very untrustworthy?

0.99 0.96 1 0.91 0.83 0.81 0.76 0.8

0.6 0.35 0.4 0.24 0.24 0.23 0.2 0.11 0.05 0.07 0.01 0.02 0.03 0 Trustworthy Untrustworthy TV Newspaper Newspaper Google maps Google maps Google search Google search Religious leaders leaders Religious Family and friends Family Mobile phone calls Mobile phone

Local government officials government Local

8 Crisis Communication Channels: Google Recommendations Chart 3: Useful Sources of Aid Information during Flood Q43a: From which sources did you receive information about aid during the flood?

100% 79% 80%

60% 47% 43% 43% 40% 29% 22% 14% 20% 11% 9% 3% 2% 1% 0% TV Volunteers Volunteers Neighbours Neighbours BPBD Staff/ BPBD officials leaders leaders Google Maps Google Maps BPBD Hotline volunteers volunteers Google Search Google Search From community From Local government government Local Other community Other Family and friends Family Mobile phone calls Mobile phone Mobile phone SMS

Table 1: Did you use any Google tools to find 4.3. Information Sources during information about the floods? the flood: Where Google tools fit Q43c – Q43e – Google tool used Google tool that 4.3.1 Google Tools accessed (N=14) yielded most During the floods, around five per cent of respondents — that Number of useful information is, a total of 14 people — accessed any given Google tool, with responses (N=14) the breakdown as follows: Number of responses Less than two percent of total respondents said they got infor- Google Search 14 12 mation that could help them from Google Search, the most accessed tool; fewer identified Google Maps as giving them Google Maps 3 0 information that they then used. (All of those who used Google Google Crisis 0 0 tools accessed Google search; some also used other Google Map tools as well.) However affected residents also travelled little, You Tube 4 0 either because the floods curtailed movement, or because they 1 0 preferred to stay close to monitor their homes; Crisis Maps, which in Jakarta’s case particularly showed what routes were Lainnya Other 3 2 flood-affected for those travelling from one area to another, are not designed to meet the information needs of those stay- As the table above also shows, of the 14 people from flood- ing within their own neighborhood. affected communities who used Google tools, 12 reported this information proved useful to source aid. This is far below other sources of information, in particular local government officials (who were themselves often responsible for distributing aid), with TV and family, friends, and neighbors following. The num- ber of respondents who used Google tools means caution is

Crisis Communication Channels: Google Recommendations 9 needed when drawing any broader conclusions; however in extremely high.3 pure percentage terms those who used both Search and Maps, Some reasons for low levels of access are clear for the most and were successful in finding useful information, is high. The affected citizens. Although crowded, Jakarta’s communications issue first and foremost is access to the tools themselves, not infrastructure for the most vulnerable is also precarious, with necessarily their utility. survey data confirming widespread electricity blackouts cutting The exception to this limited use in Jakarta is at the top level off all media and telephony services for around half of those of crisis response institutions, such as the National Disaster directly flood-affected.4 Management Agency (BNPB, Badan Nasional Penanggulangan Internews is not privy to the number of hits various Google and Bencana), UNOCHA, and Red Cross / Red Crescent (PMI, Palang Google Crisis Response services may have received during the Merah Indonesia). The Local Disaster Management Agency Jakarta floods, but given Google requested direct interrogation (BPBD, Badan Penanggulangan Bencana Daerah) used online of how these same tools were used, assumes said use was of a tools to a lesser degree. However awareness among these meaningful level. If this is the case, the most likely explanation circles does not translate to widespread utilisation of these is that those accessing the services are from the more afflu- tools by affected populations during a disaster, as shown by ent, and less vulnerable, urban middle class. Internews believes the survey numbers. this is a safe assumption, although further research would be required to definitively confirm it. 4.3.2 Rokatenda and Aceh case-studies The survey only targeted flood-affected in Jakarta. Google tools While this may explain the lack of access to Google Crisis were not utilised by affected communities in the other two Response tools (and Google services in general) by those resi- case studies, Rokatenda and Aceh. In Rokatenda, this can be dents who could benefit from them most, it does not explain attributed to a lack of infrastructure and connectivity for the their apparent low level of utilisation by the bulk of responder affected populations themselves. In Aceh, this can be attrib- organisations as indicated through the key interviews. Many uted to the short time-frame of the earthquake and tsunami organisations, including civil society, government, and vol- warning, with its accompanying power blackout and GSM con- unteer networks, have extensive experience and practice in gestion also rendering both desktop and mobile access largely disaster response, and it is somewhat unexpected that the new impossible. generation of ICT tools is not utilised to a greater extent. While some institutions in Jakarta used Google tools to a All of the examples Internews directly encountered are greater degree than the affected population, this also was described below. Observations drawn both from these specific much reduced in the Rokatenda and Aceh case studies. In examples, and from the broader appreciation of communica- Rokatenda, this can be attributed to a combination of lack of tion ecosystems underpinning the research, are then used to awareness, and reduced connectivity. In Aceh, it can be attrib- inform a range of suggested possible initiatives. uted to the same power cuts and GSM congestion that affected the general population. (For examples where Google tools were used in these case studies, see Sections 5.3, 5.4, and 5.6 of these Recommendations.)

4.3.3 Use of Google Tools in the Jakarta Floods: Some Observations The low levels of Google use, and broad absence of Crisis Response tools utilized during the Jakarta floods by both affected communities and response organizations was some- what unexpected. Jakarta is Indonesia’s capital, and it’s media and communication infrastructure — while sometimes over- crowded — is the most extensive in the country. Technology has a high profile among consumers, with Google’s leading ser- vices including search and maps well-known, and gaining mar- ket share on mobile platforms with the fast take-up of Android- 3 For more detail, see Crisis Communication Channels, 2.3 powered handsets; social media participation numbers are also 4 Ibid, 3.5

10 Crisis Communication Channels: Google Recommendations Google Services in Disaster Response: Case Study Examples

5 5 5.1 Disaster Agencies: Jakarta The national disaster management agency BNPB plays a criti- cal role both in the Jakarta case study, and in supporting and building up the capacity of local and provincial disaster man- agement offices around the country. The agency uses Google Documents, in particular, extensively; however it has not used nor promoted Google Crisis Response tools. Data collated by the BNPB formed the basis for the Crisis Response map; how- ever due to institutional delays in utilising the data, they were usually uploaded approximately six hours later — a significant limitation in a situation that changed by the hour or faster. Many of the Agency’s other office systems are still being devel- oped, both for specific functions (such as flood prediction) and more generally. This includes, for example, systems to ensure the office itself is disaster-resilient and can keep functioning once a Children in Jakarta’s January 2013 floods. Photo by Febi Dwirahmadi. crisis strikes. As they are put in place it may be possible to incor- porate a wider range of services including from Crisis Response; however at this stage that possibility is somewhat speculative. humanitarian sector.6 Jakarta’s local disaster management agency, BPBD, reportedly As the Jakarta case study shows, for agencies with intensive asked for greater deployment of a range of Google tools prior to coordination functions there remain significant bottlenecks in the floods, to assist with better data flow. However as related by data flow. This both affects institutional operations themselves, one source, the response was that the tools are too expensive for and the ability of outside actors – ranging from media to vol- this purpose, and so adoption was not pursued. At this point such unteers to, for example, mobile developers willing to produce information is hearsay, but it bears further investigation. applications for disaster response – to use emergency data in The Governor’s Office utilised Google Maps in the early stages of the most effective ways possible. This means that while for the floods, mapping reports received via SMS to their location, example BNPB data gain in functionality through using Google and logged directly using Android phones with geo-tagging. tools, other factors impose delays and limitations on this The office’s use of digital tools evolved as the floods unfolded, potential. developing some practices and understanding that allow for Disaster agencies in the other case studies in Aceh and further development. This included incorporating Bukapeta and Rokatenda / Sikka did not report using Google tools to any Mapbox in the later stages of the disaster. significant degree (Sikka’s BPBD7 did not use any online tools, UNOCHA’s Jakarta office supported and disseminated aware- inclusive of email). ness of many of these initiatives. This included circulating details of the Google Crisis Response tools throughout the 6 See http://indonesia.humanitarianresponse.info/news/google-crisis- response-resources-related-2013-jakarta-flood (accessed June 07, 2013) 5 For more detail, see Crisis Communication Channels, 3.6 7 For more detail, see Crisis Communication Channels, 4.5

Crisis Communication Channels: Google Recommendations 11 8 One of these initiatives has been to plot the locations of their 5.2 Media: Metro TV, others volunteer members on Google Maps. This would enable them By and large even the largest, most sophisticated players made to directly identify and contact a given volunteer to find out and little use of specialised digital response tools in the case study share what was happening in their area in any future disaster, a crisis reporting. Metro TV was the only outlet interviewed that capacity they currently don’t have. As a digital platform, Google utilised Google Maps. They took screen-shots for news report Map allows RAPI far more flexibility in plotting and tracking its background graphics. This under- utilisation is likely driven by a members and the information sources they represent, which range of factors, including skill levels and awareness: a media can enable faster sourcing of targeted information from where news narrative that emphasises government failure or sensa- those members are located in the event of a disaster. However tional scandal also potentially does not prioritise new practices. while plotting locations online gives great flexibility, any such A couple of print publications had greater engagement and map would still need to be downloaded and printed, given the awareness. Kompas did publicise both My Map9 and Crisis Map10 risk or even likelihood of connectivity breakdowns and power flood resources. However while it is the leading print outlet with blackouts at the moment of need.12 a strong online presence, it was not a significant source for those RAPI isn’t the only one considering such moves. One failure most affected by the floods, and as the research shows this cov- of the April 2012 warning was the lack of clearly-identified erage did not result in uptake of the tools for those communities evacuation locations and routes. Aceh’s Tsunami and Disaster — although it may have helped promote the tool’s use among Mitigation Research Centre TDMRC has begun producing more others with broader online access. However the paper’s engage- detailed maps to support better disaster-preparedness mea- ment and interest may be something to build on. sures, including pinpointing and possibly developing evacua- A few journalists from other outlets also used Google Maps as tion locations, using satellite data from Quickbird. All of these a general navigational tool, although this was reportedly some- efforts naturally cannot be fully effective on their own; mapped what limited as they already knew their targeted areas. Beyond evacuation points also require clear signs on the ground, and so this, a number of media outlets also have Android applications on. But they can fill an important, neglected gap. for audience access. 5.4 Coordination: Google groups 5.3 NGOs & Civil Society: RAPI11 mailing list The volunteer association of hand-held radio users RAPI The Milis Bencana Google group and its mailing list, ben- (Radio Antara Penduduk Indonesia, sometimes translated as [email protected], is a significant resource for disaster Indonesian People’s Radio Network) was the single most resil- responders across the country, providing disaster responders ient form of communication in the Aceh earthquake and tsu- with both updates and access to information and support; it nami warning in April 2012, meaning it did not shut down with has over 4,400 members. (As one of many examples, Caritas power blackouts, or suffer congestion as did the GSM networks, used it to gain technical information on how to purify water and had practitioners who were well-integrated into various in the Rokatenda response.) On the technology front, for communication flows. It provided vital information to an infor- example, it allows responders to connect, to support provided mal network of local officials, responders, media, and residents. by individuals and groups such as Air Putih, and its Ushahidi- Staunchly analogue, since that event they have begun explor- powered media centre.13 ing ways in which they can prepare to respond faster and better in the future.

8 For more detail, see Crisis Communication Channels, 3.7 12 RAPI isn’t the only one considering such moves. One failure of the April 2012 warning was the lack of clearly-identified evacuation locations 9 See http://tekno.kompas.com/read/2013/01/17/1157270/Pantau. and routes. Aceh’s Tsunami and Disaster Mitigation Research Centre Banjir.Jakarta.dari.Google.Maps (accessed 07 June 2013). However TDMRC has begun producing more detailed maps to develop better several observers stated that the data on this map was frequently disaster-preparedness measures, including pinpointing and possibly inaccurate or out-of-date when it was posted, and so became targeted developing evacuation locations, using GIS data from Skybird. All of for criticism. these efforts naturally cannot be fully effective on their own; mapped 10 See http://tekno.kompas.com/read/2013/01/18/11044499/Inilah. evacuation points also require clear signs on the ground, and so on. But Versi.Resmi.Peta.Banjir.Jakarta.dari.Google (accessed 07 June 2013) they can fill an important, neglected gap. 11 Ibid, 5.8 13 See http://www.mediacenter.or.id/ (accessed 07 June 2013)

12 Crisis Communication Channels: Google Recommendations 5.5 Digital Volunteers: ODOS Indonesia has developed an active network of volunteer digital responders, whose roles in several disasters has been nothing short of ground-breaking. An extended community of practice has emerged, with participants in one disaster able to access different levels of support and suggestions from those with previous experience. ODOS (One Day One Service, a short- term volunteer mobilisation effort; seeCrisis Communication Channels, Section 3.11) in Jakarta,14 and Flobamora Bloggers in the Rokatenda response,15 are two case study examples. As described in the main report, ODOS did use Google Maps and , but in a piecemeal or haphazard way, together with a collection of other tools. Use was limited to volunteers occasionally using Maps on Android phones to find directions; and mark coordinates locating displaced residents — but these were then mapped manually to an Ushahidi map. Google Docs were used for some data on volunteers, donations, and 5.6 Digital volunteers: so on, although they were then publically displayed as screen- Blogger groups shots transferred to a Wordpress blog. As mentioned in the Crisis Communication Channels report There are a variety of reasons for this. One is the organic rapid (section 4.1), blogging generates an ‘infrastructure’ or network assembly of tools by volunteer groups at the time of the response. for online participation across the country, with many bloggers However a potentially more fundamental question is why the forming local groups and engaging with issues in their area. most skilled and leading digital volunteers are not using Google They are often among the communities that directly respond Maps along with other tools more extensively in the first place. to disasters, as Flore’s Flobamora community demonstrates Several developers involved in crisis response efforts — ranging in the Rokatenda case study. One of the main platforms used from freelancers to NGO staff to private sector mobile app devel- is Blogger.com, which was the fourth most visited site in the opers — cited cost structure as a major reason: charges for use country in 2011.16 (Some blogs were also established in Jakarta in apps of Google Maps in particular was beyond their capacity specifically in response to the floods, such as ODOS’ own; many and indeed ran counter to their goals. (Internews received dif- of these also reportedly used Blogger.) ferent explanations of the charging structure, but understands it These are the examples encountered across Crisis links cost to the number of hits. An application with a humanitar- Communication Channels’ case studies. There are of course ian goal wants as many hits as possible, but requires them to be a number of other initiatives outside the case studies them- unpaid to promote access.) selves, including by Yogyakarta’s Universitas Gajah Mada (UGM) As a result, several developers were moving towards Open to trial means of crowdsourcing humanitarian needs using Street Maps for such functions instead, along with the already- Google Maps,17 building on the 2010 Merapi volcano eruption established Ushahidi platform. It is this regular, if not daily, experience. However the case study examples, including the engagement with a given platform that provides the familiarity data from the Jakarta neighbourhood survey, gives a detailed and awareness that makes it useful at the time of a disaster. illustration of the use of both general Google and dedicated Without greater knowledge of Google’s market strategy, at Crisis Response tools in disasters; and of several of the factors face value this appears to be an example in which business influencing greater adoption and utilisation. decisions surrounding a platform undermines that platform’s potential in crisis response. Whether possible means exist to circumvent such a Catch-22 is an internal Google question and beyond our capacity to suggest. 16 Lim, M., 2011. @crossroads: Democratization & Corporatization of Media in Indonesia. Participatory Media Lab Arizona State University, and Ford Foundation. Available at: http://participatorymedia.lab.asu. 14 Ibid, 3.11 edu//Lim_Media_Ford_2011.pdf (accessed 23 May 2013). 15 Ibid, 4.10 17 See http://merapi-partisipasi.ugm.ac.id/ (accessed 07 June 2013)

Crisis Communication Channels: Google Recommendations 13 Potential Initiatives

6 A number of initiatives can boost the use of online tools and Google services in disaster response. Many are informed by the factors highlighted in the Crisis Communications Channels report. Others respond to more general trends in Indonesia’s market, and Google’s engagement in particular.

Crisis Response itself.) The more actors are familiar with Google Maps, , and so on in daily life, the more likely they are to turn to these and closely-associated tools in a disaster response. However, even as expansion strategies increase the potential uptake of crisis tools and lay the foundation for future use, this does not mean Crisis Response tools reach their potentialAs suggested in Communication Crisis Channels. Natural mar- ket expansion among populations already familiar with what Google has to offer does not enable the most vulnerable popu- lations to manage better during a crisis. As Communication Crisis Channels explores in detail, gaps are not only due to sim- ple lack of access by individuals or communities. Gaps occur because the ‘links in the chain’ of communication ecologies are often fragmented and poorly-aligned to meet urgent needs. Palue residents displaced to host family houses in Ende regency, Just because a particular tool or service is taken up in one sec- Flores said the island was their home, and the place of their an- tor of this ecology, or even by a few actors in one sector, does cestors. But remembering local history going back several decades, not mean that the information it can provide will flow through they feared eruptions might continue for another three years. They were afraid for their relatives still in their home village, but there to others and deliver to those who need it. was nowhere else for them to stay. Photo by Matt Abud / Internews. Communication ecosystems go beyond the need to provide ready-made tools and services. Instead they highlight the The strongest influence driving expansion and adoption of relationships and methods needed to make sure those tools Google Crisis Response tools is Google’s own growing presence work. A service that gathers and presents emergency needs in the Indonesian market. The use of these tools — or indeed data perfectly, even if disseminated to all the emergency data- of any communication channel — is underpinned by familiar- gathering institutions (something which still has a long way to ity and engagement built up over an extensive period before a go), will still not fulfil its function if the institutions charged disaster occurs. In times of crisis, people overwhelmingly turn with gathering that data do not share the same understanding, to what they already know; while there may be some adoption practise together, and collaborate to provide the information of new practices, this is unlikely to occur with a wide num- needed. Communities with established media consumption ber of people, especially in emergencies occurring over a short habits will not turn to new sources unless prompted by those time-frame. (Crises that extend over months or years may be they already trust; and so on. Developing or adapting a tool or different, but these are not the contexts targeted by Google service in and of itself is not a recipe for success.

14 Crisis Communication Channels: Google Recommendations For Crisis Response tools and services to best meet potential, effective adoption within vulnerable communities also calls their implementation needs to leverage the synergies and fos- for community development approaches. This is not typically ter or build on the relationships that drive information flows a skill-set towards which companies are geared; efforts along through communication ecosystems. This can take place at a these lines therefore require appropriate partnerships that can greater or lesser depth. It means learning from the experience bring community development approaches into the process. of others as they have gone down the same path – for example, the work of AIFDR and Open Street Maps18 in building disaster Possible Initiatives: mapping capacity at the local government level; or World Bank n Initiatives seriously targeting ‘last mile’ neighbourhood initiatives promoting better data capacity among key govern- social networks will necessarily be involved, and require ment response institutions;19 or drawing on general lessons partnerships that can test and pilot approaches. This may from the ground-breaking experiences of Jalin Merapi’s 2010 well be beyond the remit and operations of Google Crisis volcano response.20 Response. However it would be disingenuous to propose But beyond that, it means establishing partnerships that can a simple ‘one-step’ improvement — a particular tool or help place Crisis Response tools among the channels vulner- device — that could definitively improve impact here, given able communities and disaster response organizations already the inevitable diversity and unpredictable factors at work. 21 use. This will help promote the effective use of and familiar- Possibilities therefore include: ity with those tools before a disaster takes place. Given the n Ensure tools are engineered to be accessible to the most various influences at work in communications ecologies, this is vulnerable. In particular, the Crisis Map is not optimised most likely useful through targeted small pilots before expand- for mobile access, nor as an app. Yet in Indonesia, mobile ing to wider dissemination, methodologies common in institu- access is king, particularly at the moment of a disaster tional and community development settings. when internet cafés are far less practical. (Re-designing and testing such tools can be part of or incorporated into the The potential initiatives outlined below follow this approach, next point, pilot projects, below.) grounded in Internews’ understanding of the challenges, learn- ing processes, and long-term hands-on experience of what n Partnership with local NGOs to pilot local projects using works in such contexts. Each set of initiatives is preceded by digital tools to cross ‘last mile’ communication barriers. This a brief contextual outline, drawing on the examples of existing would include the design phase of the project, including any Google tool use above; the themes and features of the overall adaptation of particular tools (for example to feature phone Crisis Communication Channel report (including relevant page requirements). Candidate partners may include technology references for more detail); and observations raised by inter- NGOs such as Air Putih or ICT Watch; or community develop- 22 locutors on relevant existing Google strategies in the country. ment NGOs such as World Vision International or PMI. n Develop training materials on use of crisis response tools, including YouTube videos, in local language with accessible 6.1 Neighbourhood and attractive local case studies, targeting both community Social Networks members and community development workers. n Support — whether technically, financially, or in public or Each community is complex. Access to communication, informa- private forums — advocacy efforts that seek to promote tion flows, trust, and translating communication into action usu- connectivity policy that will enhance last-mile communica- ally involve multi-layered dynamics, especially in locations with tions. This may include support for business lobbies that aim significant digital — and social — divides as in Indonesia. But for greater clarity on spectrum regulation or ISP provision; this is the essential ‘last mile’ where communication received and shared by individuals can make the most difference. 21 Many possible initiatives have repeat iterations under the categories At the mass scale, existing market strategies aimed towards below. For example, localised training materials appear under several categories including Neighbourhood Social Networks, Digital Volunteers; promoting tool adoption are the single biggest factor. But and NGOs and Civil Society. This is not repetition for repetition’s sake; again with the example of training, materials that target only 18 See http://hot.openstreetmap.org/updates/2011-07-03_hot_in_ Neighbourhood Social Networks would necessarily differ from material indonesia (accessed 07 June 2013) that also targeted additional areas. 19 See https://www.gfdrr.org/ (accessed 07 June 2013) 22 WVI is mentioned here as they have engaged in community flood 20 For a summary of Jalin Merapi’s efforts, see http://www.amarc.org/ preparedness in Cawang, Jakarta, with some successes and some documents/Caribbean_Conference/CR_ResponseJALIN_MerapiEruption_ lessons learned as highlighted in the study. See Crisis Communication EN.pdf (accessed 23 May 2013). Channels 3.4.1, 3.4.3

Crisis Communication Channels: Google Recommendations 15 Aceh 2004 tsunami damage. Many communities highlighted communication as a major concern. Warnings were poorly understood, unclear, late, or did not come at all. When warnings did come, many were confused regarding their meaning, and about what action to take. Photo by Oren Murphy

or support for less intrusive surveillance as proposed in list of potential tools for each one of these sectors. The point, upcoming ‘convergence’ regulation. These issues will drive however, is to establish the relationships that allow the tool to the nature of connectivity in the future for Indonesia, be developed, or calibrated, and trialled, in a specific context including whether it meets last mile requirements as much that integrates within communication ecosystem dynamics. as it could; Google’s high profile gives it an immediate plat- *Note: Many possible initiatives have repeat iterations under form. Any potential advocacy support must necessarily be the categories below. For example, localised training materi- congruent with business strategies in the country. als appear under several categories including Neighbourhood The processes and relationships outlined here may result in Social Networks, Digital Volunteers; and NGOs and Civil Society. new or re-calibrated tools that target neighbourhood social This is not repetition for repetition’s sake; again with the exam- networks. These may, for example, integrate with neighbour- ple of training, materials that target only Neighbourhood Social hood-level SMS ‘trees’, used by some local Indonesian Red Networks would necessarily differ from material that also tar- 23 Crescent operations; or by drawing on models of ‘local’ or geted additional areas. ‘bound’ social networks (international examples being Path or Nextdoor). They may include tools that enhance or connect to local volunteer communication associations such as RAPI or 6.2 Government Agencies. local community radios; or means of better leveraging the con- Whether key government agencies use online data tools, solidated penetration of feature phones. including Google tools, is one issue; whether they are set up However for neighbourhood social networks, and for the other to use them effectively is another, as shown by the informa- sectors noted below, the technicalities of the tool or service tion flow between in particular the BNPB (national) and BPBD are not the point at this stage. It is possible to compile a long (local) in the Jakarta floods. It is beyond Internews’ brief to engage in deep institutional 23 Crisis Communication Channels 3.9

16 Crisis Communication Channels: Google Recommendations analysis of BNPB or other agencies, as it is beyond the goals of Google Crisis Response to engage in direct institutional develop- ment. Yet improvements in a key institution’s data capacity can have many flow-on effects. Faster availability of BPBD data would have cut delays in uploading to the Google Crisis Map; if such data were available it would also enable a range of secondary actions including for example apps by independent mobile devel- opers, and access by groups ranging from ODOS to PMI. Initiatives can range from practical engagement or accompani- ment; or simply by making sure the capacity of key national institutions in a response are fully-understood, as a first step towards appreciating what practical applications of digital crisis tools can be achieved.24 In the end, this is not just about what tools are used, but what ongoing relationships and coopera- tion is needed to make sure data flows through and is used effectively (for example, the relationships between BPBD and the various government departments such as health and police that provide initial reports must be functional before a disaster hits, not merely established or revived at that moment). need first to be aware themselves of what is possible. While Possible initiatives: outlets such as Kompas have covered and so promoted aware- n Develop training modules targeting the needs of staff in ness of some tools, their audience does not include the most Government agencies. vulnerable who typically access popular TV channels above all n Draw on lessons of others involved in supporting govern- else. Buy-in by local media to both use, and promote, crisis ment disaster response capacity — such as AIFDR,25 the communication tools has the potential to fuel a major uptick World Bank, or UNOCHA. This may inform further individual in awareness, incorporation into preparedness, and use of such initiatives on the part of Google; or it may point to ways in tools in time of need. which Google can integrate with and support existing initia- There are challenges to this — conventional media, especially tives currently being implemented by these partners. the most profitable, by and large follow the immediate news n Review existing Google business operations in Indonesia to agenda and do not engage in long-term reporting or program- see if calibration to better meet humanitarian needs is pos- ming in support of behaviour change, a real challenge for disas- sible (see Google Business Strategies, below). ter preparedness.26 However new technology, and humanitarian needs, have enor- 6.3 Media mous appeal among many audiences and consumers. This means a level of media buy-in to promoting awareness of digi- Conventional media maintains most if not all of its power to tal crisis tools has more chances of success than many other drive discussion and generate awareness across the country. more ‘mundane’ disaster preparedness measures. It is a potentially powerful way to disseminate awareness and promote use of new crisis communication tools among the public — but as the Metro TV example shows, media outlets

24 For example, the BNPB as part of its institutional development is considering establishing specialized mapping rooms for emergency control centers, a potential connection-point for some Google Crisis Response services. Accompanying this development process, even on an occasional basis, can help identify whether such connections are useful as they emerge. 26 There are some exceptions to this. For example, interviewees cited 25 For example, the AIFDR is already reviewing and gathering lessons the efforts by Kompas newspaper, as well as several other local media learned from the data management experiences during the Jakarta including the radio station Rumoh PMI FM, for efforts to review and floods across the main government agencies promote disaster preparedness in their reporting.

Crisis Communication Channels: Google Recommendations 17 Possible initiatives: Nevertheless the eventual, inevitable rollout of more advanced n Develop training materials outlining how journalists and edi- technologies, smooth or otherwise, can only help accelerate tors can leverage Google tools for better reporting the take-up of more sophisticated devices and the expand- n Target increased media coverage of the potential and uses ing mobile data sector. This represents opportunities, as does of Google tools in a disaster response. This can be through Android’s rapidly-increasing share of the market; however the direct advertising, or by using a strategy that encourages mix of technologies across the digital divide, in particular the greater media news coverage of the same. (See the follow- expected ongoing prevalence of basic feature phones for the ing two points for possible means of achieving this.) most vulnerable, also needs to be taken on board in any com- prehensive initiatives.30 n Promote prizes, for example for best use of technology in humanitarian reporting Possible Initiatives: n Highlight local case studies of technology for humanitarian n Co-sponsor with telecommunication companies initiatives response in (some of which has already been noted above, such as ‘lessons learned’ sessions, humanitar- done, for example a Google ad two years ago featured the ian hackathons, and competitions. Blood for Life initiative27), with links pointing to training and n Sponsor humanitarian mobile application design using case study material. Google tools, and data made available by those tools.31 n Ensure Google Crisis Response spokespersons are avail- n Partner with telecommunications providers to test, and if able and reach out to leading media outlets at the time of necessary design and adapt, apps or services that can effec- a disaster. This includes press releases, any compilation of tively operate in a range of contexts across the country. citizen video that is appropriate, and so on. Providing a news This may include, for example, providing services in low- story at the time of need is the best way to build awareness bandwidth areas using feature phones; identifying where of available tools and resources that can most help. less-popular but more resilient CDMA phones can serve as Several of these initiatives may require a local partner, whether a critical bridge to a vital area, and so on. In one illustration, an NGO, professional media association, or other. CDMA phones with effective feature-phone data delivery could potentially serve as a bridge to local media or net- works like RAPI in Aceh, ensuring data availability connects 6.4 Telecommunications Sector to the essential communication resilience and redundancy Ongoing developments in Indonesia’s mobile telecommunica- that these entities provide.32 tions sector, and moves towards providing greater connectivity around the country, will create major shifts in the landscape into the future — but exactly how is difficult to predict, given 6.5 NGOs & Civil Society the mix of policy, political, economic, and legal factors at Indonesia’s NGOs and civil society are, typically, a mixed bag 28 work. Some telecommunications companies, or their local when it comes to adoption of technology, including in crisis offices, are more involved in providing crisis response support response. As outlined in Internews’ New Digital Nation report,33 than others; some initiatives have moved forward, while others there is generally a significant disconnect between national or 29 are somewhat in limbo. local Indonesian NGOs and the developer community, inhibit- ing the uptake and adaptation of digital tools. For smaller NGOs, this is often due to lack of awareness, time, or access 27 See https://www.youtube.com/watch?v=XkUtEJtV46U (accessed 07 June 2013) to those with skills. For larger NGOs, it can also be an issue of 28 For more detail, see Crisis Communication Channels, 2.2 and 2.4 29 One positive example is the ongoing manner in which Telkomsel in Aceh has developed its response operations, and Indosat’s CSR efforts – although in Indosat’s case, initiatives to provide emergency communications are vetted through its marketing department, thus likely mixing mandates and imperatives and creating bureaucratic delays. One example that has been held up is efforts by PMI and IFRC 30 For more detail, see Crisis Communication Channels, 2.2 to introduce the emergency SMS gateway TERA to the country (see 31 For example, Indosat already sponsors app development for http://www.ifrc.org/en/what-we-do/beneficiary-communications/ educational purposes through its CSR program. tera/); unresolved issues surrounding who would ‘control the button’ to send out – PMI or government – along with complications 32 Crisis Communication Channels, 5.8 surrounding the fragmented telecommunications sector itself, have 33 For more detail, see https://www.internews.org/research-publications/ slowed progress on this. indonesia-new-digital-nation (accessed 07 June 2013)

18 Crisis Communication Channels: Google Recommendations bureaucracy that makes replication of practice difficult.34 on. Coordination forums are themselves key mechanisms to However as noted under above Neighbourhood Social Networks disseminate awareness of tools and any training materials, as and elsewhere, civil society and NGOs have extensive engage- well as other activities noted above, including prizes and so on. ment and expertise in many other areas essential to ensuring Coordination forums can also serve to inform any locally-tar- digital crisis response tools meet their potential, including geted activities, including for Neighbourhood Social Networks. community development, some cases of institutional develop- ment, and so on. If effective ways of engaging with actors in this sector can be found without diverging from existing opera- 6.7 Digital Volunteers & tions and mandates of Google Crisis Response, this may lead to Communities of Practice examples of practice that can be disseminated and replicated Volunteer digital responders are a leading community of prac- elsewhere. tice for new technology tools, but — as demonstrated by ODOS and on a smaller scale Flobamora — they are extremely Possible initiatives: organic in their genesis and the process whereby skills and n Target partnerships for Neighbourhood Social Networks pur- systems are learned and disseminated. For example, the key poses, and potentially in support of Digital Volunteers; see individuals driving ODOS had prior experience in the 2010 respective notes above on each of these areas. Merapi response, which has become the benchmark of digital n Develop training materials targeted civil society and NGO mobilisation in crisis response within Indonesia. Several of the staff, incorporating local context as noted above. individuals involved in such efforts have high profile in blog- ger communities; Flobamora volunteer bloggers reached out to these same individuals when they began their effort. Sharing 6.6 Coordination and support is a big part of such networks’ strength, with While not a direct ‘actor’, as demonstrated throughout the examples of practice laid down in one response then informing Crisis Communication Channels’ case studies, coordination in the next. However this also means that sub-optimal practices a humanitarian response is critical, including for communi- can be improvised — and then replicated in future responses, cation and the roles of platforms such as those provided by simply because they are what was used before. Google within this. Such coordination is far from Google Crisis The role of such networks need therefore to be appreciated as Response’s function. However an appreciation of its impor- ‘catalysts’ for an emerging set of practices, as well as for their tance may help inform some other initiatives. direct contributions. Possible initiatives: Volunteer digital responders form loose, organic, and often temporary networks. However they regularly draw on or are n Provide a direct Google contact and spokesperson during a facilitated by a small number of key players, several of whom disaster response, who is able to demonstrate the use of play roles in related NGOs including ICT Watch, Air Putih, and appropriate tools, deliver and / or point towards rapid train- Combine Resource Institute.35 This means it is possible to posi- ing materials, and respond to media and other PR enquiries tively influence emerging practice through targeted engage- (also noted under ‘Media’, above). This may be a Google ment with a few practitioners, notwithstanding the fluid nature Crisis Response staff person, or it may be a representative of the networks themselves. from a local organization or network who has been briefed and / or trained, and brought into the role. Any positive evolutions in practice developed in this way would These include, again, training materials that highlight how par- then be positioned to disseminate to other actors in humani- ticular tools and services can support coordination efforts and tarian response over time, as digital technology and its human- why they’re important — such as open data formats and so itarian uses become more appreciated and widespread.

34 For example, some Red Cross / Red Crescent societies developed Further Communities of Practice groundbreaking communication practices across a range of platforms in the response to the 2004 tsunami in Aceh. However a number Beyond digital responders themselves, Indonesia has a wide of IFRC members noted that many of these practices are only now range of developer and consumer user groups, with many local being adopted by PMI and other societies operating in Indonesia, due to difficulties in ensuring enough people across the organization ‘chapters’, and frequently high levels of enthusiasm. These can both appreciate the importance of such measures, including digital communication tools; and to commit to the extent necessary to get 35 Internews understands Google already holds relationships with some if organizational buy-in to take them forward in other contexts. not all of these examples.

Crisis Communication Channels: Google Recommendations 19 range from Android user groups to Google Developer Groups, Red Hat developers, and more. They represent real potential to engage in and promote digital platforms and tools.

Possible Initiatives: n Support or facilitate ‘lessons learned’ sessions with digi- tal volunteers after significant disaster deployments. This includes capturing the ‘lessons learned’ for subsequent dissemination. n In partnership with leading digital responders — either organisations or influential individuals — develop or adapt a suite of tools responsive to the range of past and likely future disasters experienced by the country. This includes ‘lessons learned’ sessions after the suite has been deployed. This may include local hackathon sessions and so on, depending on what is practical with the skills present in country. n Develop training materials, again including video, and in local language using local case studies, explaining the suite of tools and their deployment, targeting digital volunteers. n Foster further developer and user group networks with a humanitarian focus. To some degree this already exists, but the prevalence of Android user groups, as well as specific Google developer groups and business groups, suggests this can be expanded far more with some targeted effort. This may include regular G+ hangouts, Facebook groups, and further Google group email lists, that foster interaction. Methods can include events or gatherings; prizes for a range of activities (best proposed tool adaptation, most active local group, etc.), and so on. n Review existing Google business operations in Indonesia to see if calibration to better meet humanitarian needs is pos- sible (see Google Business Strategies, below).

20 Crisis Communication Channels: Google Recommendations List of Acronyms 7 BBM BlackBerry Messenger BNPB National Disaster Management Agency (Badan Nasional Penanggulangan Bencana) BPBD Local Disaster Management Agency (Badan Penanggulangan Bencana Daerah) CDMA Code Division Multiple Access DRR Disaster Risk Reduction GIS Geographic Information System GSM Global System for Mobile Communications HT ‘Handie-talkie’ or hand-held radio IFRC International Federation of Red Cross and Red Crescent Societies NGO Non Government Organization ODOS One Day One Service PMI Indonesia Red Crescent (Palang Merah Indonesia) RAPI Indonesia People’s Radio Network (Radio Antar Penduduk Indonesia) RT ‘Neighborhood administrator’ (Rukun Tetangga) RW Administrators of a larger geographic area and level of responsibility above Rukun Tetangga (Rukun Warga) SMS Short Message Service TDMRC Tsunami and Disaster Mitigation Research Centre UNOCHA United Nations Office for the Coordination of Humanitarian Affairs

Crisis Communication Channels: Google Recommendations 21 22 Crisis Communication Channels: Google Recommendations

Internews Washington, DC Office Internews is an international non-profit organization whose mission is 1640 Rhode Island Ave. NW Suite 700 to empower local media worldwide to give people the news and infor- mation they need, the ability to connect and the means to make their Washington, DC 20036 USA voices heard. + 1 202 833 5740 Internews provides communities the resources to produce local news and information with integrity and independence. With global exper- Internews Administrative Headquarters tise and reach, Internews trains both media professionals and citizen PO Box 4448 journalists, introduces innovative media solutions, increases coverage Arcata, CA 95518 USA of vital issues and helps establish policies needed for open access to +1 707 826 2030 information. Internews programs create platforms for dialogue and enable informed www.internews.org debate, which bring about social and economic progress. E-mail: [email protected] Internews’ commitment to research and evaluation creates effective Twitter: @internews and sustainable programs, even in the most challenging environments. facebook.com/internews Formed in 1982, Internews is a 501(c)(3) organization headquartered in California. Internews has worked in more than 75 countries, and cur- rently has offices in Africa, Asia, Europe, the Middle East,L atin America and North America.